Altruism, Turnout and Strategic Voting Behavior

We consider the problem of costly voting in a game-theoretic framework where agents are altruistic. We show that if, as usual, agents’types are assumed to be independently and identically distributed, the classical theorem of Thomas R. Palfrey and Howard Rosenthal (1985) on the impossibility of large-scale turnout essentially survives, despite the introduction of altruism. We solve this problem by introducing additional uncertainty about the fraction of altruistic agents who support a given candidate. It turns out that under suitable homogeneity assumptions, this modi…ed model is asymptotically equivalent to the rule utilitarian voter model of Timothy J. Feddersen and Alvaro Sandroni (2006, 2007), which is known to be consistent with large-scale turnout and strategic behavior. However, in contrast to rule utilitarian voter models, which typically require large-scale agreements among uncoordinated agents, our model is also compatible with several forms of heterogeneity in agents’characteristics. It is hoped that this paper will help to close the discrepancy between our understanding of costless or small elections, which is largely shaped by pivotal-voter models, and that of costly, large elections. (JEL D64, D72) I am grateful to Debraj Ray who supervised this research, which also bene…ted from various discussions with Daniel Barczyk, Gokhan Buturak, Guillaume Frechette, Anna Gumen, Melis Kartal, Yilmaz Kocer, Alessandro Lizzeri, Efe A. Ok, Pietro Ortoleva, Gil Riella and Florian Semani. I owe special thanks to Timothy J. Feddersen, Srinivasa Varadhan and Atilla Y¬lmaz. All remaining errors and de…ciencies are mine.